Optical Constants of Organic Tholins Produced in a Simulated Titanian Atmosphere" from Soft X-Ray to Microwave Frequencies

ICARUS 60, 127--137 (1984)

Optical Constants of Organic Tholins Produced in a Simulated Titanian Atmosphere" From Soft X-Ray to Microwave Frequencies

B. N. KHARE AND CARL SAGAN Laboratory for Planetary Studies, Cornell University, Ithaca, New York 14853

E. T. ARAKAWA, F. SUITS, T. A. CALLCOTT, AND M. W. WILLIAMS Oak Ridge National Laboratory, Oak Ridge, Tennessee 37830

Received January 20, 1984; revised May 23, 1984

As part of a continuing series of experiments on the production of dark reddish organic solids, called tholins, by irradiation of cosmically abundant reducing gases, the synthesis from a simulated Titanian atmosphere of a tholin with a visible reflection spectrum similar to that of the high altitude aerosols responsible for the albedo and reddish color of Titan has been reported (C. Sagan and B. N. Khare, 1981, Bull. Amer. Astron. Soc. 13, 701 ; 1982, Orig. Life. 12, 280) and [C. Sagan, B. N. Khare, and J. Lewis, in press. In Saturn (M. S. Matthews and T. Gehrels, Eds.), Univ. of Arizona Press, Tucson]. The determination of the real (n) and imaginary (k) parts of the complex refractive index of thin films of such tholin prepared by continuous D.C. discharge through a 0.9 NJ0.1 CH4 gas mixture at 0.2 mb are reported. For 250 ,~ -< h -< 1000/zm, n and k have been determined from a combination of transmittance, specular reflectance, interferometric, Brewster angle, and ellipsometric polarization measurements; experimental uncertainties in n are estimated to be -+0.05, and in k -+ 30%. Values of n (=1.65) and k (=0.004 to 0.08) in the visible range are consistent with deductions made by ground-based and spacecraft observations of Titan. Maximum values of k (=0.8) are near 1000 .~, and minimum values (=4 × 10 41) are near 1.5 /xm. Many infrared absorption features are present in k(h), including the 4.6-/~m nitrile band. © 1984 Academic Press, Inc.

INTRODUCTION solids, called tholins, are produced by the irradiation of a wide variety of cosmically The Voyager 1 discovery (or confirma- abundant reducing gases (Sagan and Khare, tion) of nine simple organic molecules as 1979). The tholin produced at Cornell ex- minor constituents of the atmosphere of Ti- hibits a visible spectrum short of the onset tan (Hanel et al., 1981; Maguire et al., 1981 ; at 0.63 /zm of the CH4 Kuiper bands very Kunde et al., 1981), surmounting an unbro- similar to that observed for the Titan aero- ken dark reddish haze layer (Smith et al., sols (Khare et al., 1981; Sagan et al., 1984). 1981), tends to confirm earlier suggestions Gas chromatography/mass spectrometry of of an organic aerosol on Titan produced in pyrolyzates prepared from this tholin re- situ from the major atmospheric constitu- veals over 75 separate compounds, includ- ents (Sagan, 1971, 1973, 1974; Khare and ing many molecules of fundamental biologi- Sagan, 1973). In experiments previously re- cal significance on Earth (Khare et al., ported, we have exposed a simulated Tita- 1982, 1983). nian atmosphere [0.09 CHd0.91 N2 at 73 A determination of the complex refrac- mb] to high-frequency electrical discharge tive index of various kinds of simulated Ti- for 4 months and have produced a dark red tan tholins over as wide a frequency range solid that we propose is similar to the ob- as possible would permit comparison with served Titan aerosol (Sagan and Khare, refractive index estimates made at visible 1981, 1982; Sagan et al., 1984). Similar wavelengths from ground-based and space- 127 0019-1035/84 $3.00 Copyright © 1984 by AcademicPress, Inc. All rights of reproductionin any form reserved. 128 KHARE ET AL. craft photometric and polarimetric observa- atmosphere (Smith et al., 1982). Cosmic tions; would allow calculations on the rays and, to a lesser extent, Saturn magne- transmissivity of the Titan atmosphere to tospheric electrons deposit most of their sunlight, and therefore on the validity of energy in the Titanian lower atmosphere, at greenhouse mechanisms to explain the an altitude still inaccessible to observation small increment (-~13°C) of Titan surface (Sagan and Thompson, 1984). Thus the tho- temperature over Titan equilibrium temper- lins produced in the present study are likely ature; would enable modeling of observed to correspond to the reddish aerosol haze in albedo/color variations both in space and in the upper atmosphere of Titan. In the re- time in the aerosols of Titan; and might be gion simulated, at the base of the Titanian useful for calculations of radiative transfer thermosphere, the ambient temperatures in the infrared and microwave regimes. are =I70°K. The laboratory experiments Knowledge of the optical properties is also were carried out at room temperature, but an essential prerequisite for putting on a because the difference in effective tempera- firmer basis the idea that Titan tholin aero- ture between the energy source for molecu- sols are in fact the major chromophore in lar dissociation and the ambient tempera- the atmosphere of Titan. ture is so large, both in the laboratory and Reliable determination of the real part, n, on Titan, the choice of laboratory ambient and the imaginary part, k, of the complex temperatures may not affect the results sig- refractive index of the tholin by measuring nificantly. the reflectivity of optically thick samples A Technics Hummer V film deposition (pressed pellets) and then performing a chamber at Oak Ridge National Laboratory Kramers-Kronig analysis proved ex- was employed. The experimental setup is tremely difficult, because the surface of the shown in Fig. 1, and the details of the elec- samples were not truly specular. It became trode and field configurations are given in evident that reliable determinations of n Fig. 2. The apparatus consists essentially of and k would require transmission measure- two horizontal aluminum electrodes of di- ments, which in turn would require the ameter -~7.6 cm, separated by a vertical preparation of optically thin films of Titan distance of -~5.7 cm, enclosed coaxially tholin. within a vertical glass cylinder of internal Thin films of tholin were generated from diameter =10 cm. Clean glass or alkali a gas mixture of 0.9 N2 and 0.1 CH4 by vol- halide substrates for the tholin films were ume at a total pressure of 0.2 mb. This cor- placed in a nearly vertical position on the responds to a radial distance from the cen- lower electrode and leaning against the in- ter of Titan of about 2825 km, just at the top side wall of the cylindrical glass reaction of the main cloud deck viewed by Voyagers vessel. The N2 used was prepared from liq- l and 2, and below most of the visible aero- uid nitrogen of minimum purity 99.997%, sol haze. The solar ultraviolet flux at X < containing <3 ppm H20 and <5 ppm 02; 900 ,~, Saturn magnetospheric electrons the CH4 was chemically pure (CP) grade, and protons, solar wind electrons, and cos- 99% purity. A typical analysis of CP grade mic rays are all able to break N2 chemical methane revealed 0.6% N2, 0.5% 02, 0.2% bonds and synthesize nitrogenous organics CO and CO2, and 0.13% C2H4 as impurities. from NJCH4 atmospheres. Saturn magne- Both gases were obtained from Linde Divi- tospheric particles and interplanetary elec- sion, Union Carbide Corporation. The gas trons deposit most of their energy at radial mixture continuously flowed through the l- distances between 3200 and 3600 km from liter chamber at a rate -~0.05 cc/sec. A 15- the center of Titan (Sagan and Thompson, mA direct current electrical discharge was 1984), a region corresponding to an ob- maintained by a 200-V potential difference served extreme uv opacity in the Titanian between the electrodes. The tholins pre- OPTICAL CONSTANTS OF ORGANIC THOLINS 129

IONIZATION • CHAMBER 15ma

M'~ots~o'~ ~,~scs. ---2OO V DC. METHANE VALVE ~l'-~ ~~ PRESSUREGAUGE VALVE

~I~LVE NITROGEN t i! -- TANK GA a-.... ;I :zz:cC~" FOR OU "3{'/ MIXED GASES ~ PU~

PUMP TO NEEDLE VALVE

Fi~. 1. Experimental apparatus, incorporating Technics "Hummer V" film deposition chamber, used in preparing thin film of Titan tholin on various substrates. sumably form by quenching and reaction of and the other being air or vacuum, is given the ionization and dissociation products in by Hall and Ferguson (1955). We consider a the region of the discharge, and then diffuse wavelength region in which a thin homoge- outward, where they are deposited on the neous tholin film characterized by a thick- substrates. Most of the films were formed ness, t, and a real part, n, of the complex on glass microscope slides, but some were refractive index is transparent, interference deposited on disks of CaF2, LiF, or CsI. fringes are seen, and the normal incidence Since the substrates were of different sizes, transmittance T at wavelength X is the geometry for film formation varied. The T = [1 + 4R2(1 - Rf2) 2 sin 2 (27rnt/h)]-J, (1) thickness at each point of a given film de- pends on this geometry, on the deposition where gf is the reflectance of the tholin time and on the time-variable configuration film. Maxima in T occur at integer values of of the discharge. Thus the thickness varied 2nt/h. For wavelength regions in which across each film, and the maximum thick- such a film, deposited on a substrate of ness varied from film to film. A compara- thickness t~, is absorbing, the normal-inci- tively thick film, with maximum thickness dence transmittance is about 20/zm, was deposited in about 3 days T = (I - R)e-~tse -4~kt/x, (2) of sparking. To obtain the tholin optical constants from the soft X-ray to the micro- where R is the total normal-incidence re- wave region, independent measurements of flectance from the film-substrate system, a~ transmittance, reflectance, interference, is the absorption coefficient (units, cm -1) and polarization were made, as appropri- for the substrate, and k is the imaginary ate. part of the complex refractive index. A general transmission equation for radi- The transmittance of substrate alone is ation impinging at normal incidence on a Ts = e-'~s/s(1 - R0) 2, (3) thin, absorbing plane-parallel and homogeneous film, separating two nonabsorbing where R0 is the normal-incidence reflec- semi-infinite media, one being a substrate tance at the substrate-air interface. From 130 KHARE ET AL.

In certain wavelength ranges, T~ was mea- a L____L___ __ I sured directly. Ideally, to obtain k at a given wavelength from a measurement of the transmittance T using Eq. (4) or (5), it is necessary to know t and n independently. In our experiments, for each measurement of T, the thickness t was measured at the same location on the film. In two wavelength ranges, the visible and the far infrared, a direct measurement of n was possible. At other wavelengths an L ."8"-- .1 approximate value was first assumed in cal- culating R. Fortunately, when k is small, an independent measurement of n can be obtained, in which case the values of R0 and R may significantly affect the calculated value b ~ 7 ~-15~° v----~ ~ of k. Furthermore, n tends to vary slowly from about 1000 A to the microwave region. Thus, for wavelengths at which n is not measured directly, a satisfactory esti- mate can be made for use in the calculation of k. Over the wavelength range studied, k varies by nearly four orders of magnitude. Thus to obtain optimum sensitivity in measurements of transmittance, we must use _, ov 1½ films of different thicknesses in different wavelength ranges. The thinnest films employed, which appeared to the naked eye as FIG. 2. Ionization chamber. Slides are placed on the faint brownish coatings, had central thick- bottom electrode and touch the glass walls of the cylin- nesses of ~0.6 ~m; the thickest, -~20/xm. der. (See Fig. I.) (b) Detail of electrode. Approximate The film thickness on a given slide tended field configuration. All parts are grounded unless other- wise noted. to increase from the periphery to the center. With all controllable parameters fixed, Eqs. (2) and (3), film thicknesses and geometries still varied from case to case, because of irreproduci- k = (h/47rt) In {[T~(1 - R)/T] bilities in the discharge itself. The films. 11 - R0l 2}, (4) nevertheless, appeared to be homogeneous or, if the substrate contributes no extinc- in color. For each region of the spectrum tion, studied, a large number of samples was prepared and measured, and the results for k - (h/47rt) In [(I - R)/7]. (5) the complex refractive index presented Values of Ts, R0, and R can be calculated, here are composites of many different ob- as required, from the Fresnel equations and servations. a knowledge of the refractive indices n and Transmittance measurements were made ns of the tholin and the substrate, respec- for 0.15/xm - X _< 70/~m and for 100/xm -< tively. Values of ns for the substrates are h -< l mm. Between 0.15 and 0.5 /xm, the known from independent measurements or transmittances of films with thicknesses from the literature (Kaye and Laby, 1966). from 0.6 to I /.~m deposited on quartz, OPTICAL CONSTANTS OF ORGANIC THOLINS 131

CaF2, or LiF substrates were measured us- LiF substrates, half of the CsI substrate ing a Seya-Namioka monochromator. The was then shielded, so that no film would be substrates were small disks about 2.5 cm in deposited on it. Transmittance measure- diameter. During film deposition, half the ments were obtained using a Perkin-Elmer substrate was shielded. Transmission mea- Model 521 ir spectrometer through CsI surements were then made through the sub- alone, and through CsI plus tholin. Trans- strate alone and through the substrate plus mittance measurements for samples depos- tholin. The thickness of the tholin film ited on CsI substrates and on substrates could be estimated by counting the interfer- alone were obtained at 2/xm -< h -< 70/zm ence fringes visible using monochromatic at the NASA Goddard Space Flight Center light in going from the region of the disk (GSFC) using a Nicolet Series 8000 Fourier with no film to the region of uniform tholin transform ir spectrometer. film thickness. When all measurements of Finally, for 100 /zm -< h -< 1000 /~m, Ts 00 and T U,) had been completed on a pressed pellets were used. Tholin, depos- given film, the thickness of the film was ited on the glass wall of the reaction vessel, measured: It was scored at the point where was scraped off and compacted at 1700 bar the transmittance was measured, and the into 13-mm-diameter pellets, as described film thickness determined at the scratch elsewhere (Khare and Sagan, 1979). Three with a Zeiss light-section microscope, to pellets were employed, with thicknesses within -+0.2 p,m, or with a Sloan interfero- 0.31, 0.485, and 0.810 mm, as measured us- metric microscope, to roughly +0.02/zm. ing a micrometer caliper. Transmittance For 0.4/.tm -< h ~ 2.5/~m, tholin samples was measured at GSFC with the same spec- were deposited on glass microscope slides, trometer. placed with one long edge (-~7.6 cm) resting Transmittance in the region 70 to 100/zm on the lower electrode of the reaction ves- was interpolated, since at these wave- sel. The part of the film deposited along the lengths all pellets were too thick and all center of the slide, parallel to the lower films were too thin for reliable measure- electrode, was used for transmittance mea- ments of transmittance. However, since no surements, each film providing a range of spectral feature was found in the reflection thicknesses along its center line. For each study of the pellets in this region, the trans- chosen wavelength, T was measured at 6.4- mittance interpolation should not introduce mm intervals using a Cary Model 14PM serious error. spectrophotometer, equipped with a mov- Ellipsometric polarization measurements able rack in the sample chamber to accept a (see, e.g., Inagaki, et al., 1976) were per- microscope slide. Once a sample was in- formed on thick tholin films in the range 0.4 serted, the data taking was fully automated /zm -< h -< 2 /zm, where k is very small, for 19 wavelengths in the 0.4- to 2.5-/zm yielding values varying from n = 1.7 at 0.4 range. Once all the transmission measure- /zm to 1.6 at 2 /zm. Values of n for thin ments had been recorded, the film was tholin films on substrates of CaF2, quartz, scratched down the middle of the slide, and and glass were also measured directly by the thickness along the scratch measured. the Brewster angle method, as modified by On a typical film, measured thicknesses Abeles (1950) for thin, transparent films on ranged from -~ 1/zm at the periphery to ~20 substrates. These measurements gave n - /zm at the center of the slide. 1.63 for 0.4 /zm -< h -< 0.6/zm. Observa- For 2.5/.tm -< h -< 40/~m, CaF2, LiF, and tions of interference effects in the normal CsI disks were used as substrates. LiF and incidence T (h) for thin tholin films on glass CaF2 are opaque beyond 6 and 9/zm, re- yielded n -~ 1.7 between 1.0 and 2.3 /zm. spectively, but CsI transmits well from 2.5 The pressed pellets also showed interfer- to 50 /~m. As with the quartz, CaF2, and ence patterns in T U,) for 300 /zm -< h -< 132 KHARE ET AL.

n = (1 - R)/(1 + R - 2R 1/2 cos O) 09N2/01 CH4 THOLIN I

and

k = -2R I/2 (sin 0)/(1 + R - 2R I/2 cos 0).

Since film reflectances were measured only from 3.1 to 50 eV, it was necessary to ex- x b,m) trapolate the reflectance to lower and higher energies. From 0 to 3.1 eV, a con- FIG. 3. Normal incidence reflectance of Titan tholin measured as a function of wavelength. stant value, equal to the measured value at 3.1 eV, was assumed for R. Because the energies are so low at E < 3. I eV, the val- 1000 ~m, implying n = 2.0 in this wave- ues of k for E > 3.1 eV, obtained from the length region. Kramers-Kronig analysis, are not very sen- Reflectance at near-normal incidence, R, sitive to this assumption. (Final values of k, was obtained using tholin samples depos- derived below, exhibit significant variation ited on glass slides. Between 0.025 and 0.13 with wavelength, and the anticipated infra- /~m, a McPherson Model 247 grazing inci- red spectral features.) Above 50 eV, an ex- dence monochromator was used, while ponential extrapolation, R - R(50) from 0.13 to 0.4 p,m the Seya-Namioka exp[-0. 135 El, was assumed, where R(50) monochromator was employed. The values is the measured value at 50 eV, and E is in of R obtained are shown in Fig. 3. electron volts. The value of the exponent Analysis of the data described in the fore- was equal to the slope on a semi-log plot of going paragraphs was performed by means the measured near-normal-incidence reflec- of two separate Kramers-Kronig analyses. tance, R vs E, above =25 eV. The integra- For 250 A -< X -< 1500 ,~, the absorption in tion in Eq. (6) was not carried to infinity. the tholin was too large to obtain k from Instead, a high-energy cutoffof 175 eV was transmittance measurements; thus, the ex- chosen to make the calculated k values perimental reflectances, from 0.4 /~m (3.1 equal to the experimentally determined k eV) to 0.025/~m (50 eV), were analyzed to values at the highest energies at which k yield tholin n and k values over this wave- was measured, i.e., in the region from --0.2 length (energy) range by employing the ~tm (6.2 eV) to 0.15 gm (8.3 eV). The reflec- Kramers-Kronig relation between O(E), the tance data were also analyzed using a phase change on reflection at energy E, and power law rather than an exponential ex- the normal incidence reflectance R (see, trapolation for E > 50 eV. Values ofk in the e.g., Jahoda, 1957): region of high absorption, where k could lnR (E') not be measured directly, were found to bc O(E) = --~. (E,)2 _--~2 dE'. (6) the same with either extrapolation. The values of k from 0.025 to 0.15 /~m, In addition, obtained from the Kramers-Kronig analysis of R, were then combined with the ex- O(E) = arc tan [2k/(1 - n 2 - k2)] (7) perimentally determined k values at longer and wavelengths and the interpolated values between 70 and 100 /~m. The entire k spec- R = [(n - 1) 2 + k2]/[(n + 1) 2 + k2]. (8) trum is shown in Fig. 4, and tabulated in Thus a determination of O(E) from R by Table I. The Kramers-Kronig relation be- means of Eq. (6) allows n and k to be deter- tween n and k (see, e.g., Inagaki et al., mined from Eqs. (7) and (8); i.e., 1977), given by the dispersion relation OPTICAL CONSTANTS OF ORGANIC THOLINS 133

TABLE I

REAL (n) AND IMAGINARY (k ~- a 10 b) PARTS OF THE COMPLEX REFRACTIVE INDEX OF THOLINS MADE FROM 0.9 NJ0.1 CH4

k n ~, k (p.m) (/~m) a b a b

920.0 3.0 3 2.17 2.938 6.0 2 1.59 850.0 1.0 2 2.17 2.818 2.4 2 1.58 774.9 2.9 2 2.16 2.743 1.1 2 1.59 688.8 4.7 2 2.16 2.695 4.1 3 1.60 563.5 7.0 2 2.15 2.422 1.2 3 1.62 387.4 1.0 2.12 2.403 8.5 4 1.62 229.6 1.4 2.07 2.393 8.0 4 1.62 172.2 1.6 2.04 2.214 8.9 4 1.63 140.9 1.6 2.03 2.019 7.2 4 1.63 121.5 1.9 2.02 1.873 5.2 4 1.63 81.57 2.1 1.93 1.823 4.4 4 1.64 56.35 1.9 1.86 1.813 4.2 4 1.64 36.46 1.5 1.81 1.802 4.0 4 1.64 31.00 1.4 1.81 1.381 4.1 4 1.64 22.14 1.8 1.80 1,357 4.2 4 1.64 18.23 2.1 1.76 1.192 5.2 4 1.65 17.71 2.1 1.74 1.148 6.4 4 1.65 14.42 1.7 1.67 1.016 1.0 3 1.65 12.91 1.4 1.67 0.8731 2.4 3 1.66 11.70 9.7 2 1.64 0.6888 8.8 3 1.68 11.07 7.9 2 1.66 0.5635 2.3 2 1.70 10.51 7.5 2 1.67 0.4428 6.0 2 1.72 8.731 9.2 2 1.71 0.4133 7.6 2 1.69 7.653 1.3 1 1.72 0.3874 9.1 2 1.66 7,044 1.7 1 1.69 0.3542 1.1 1 1.63 6.666 2.2 1 1.69 0.3263 1.3 1 1.64 6.457 2.6 1 1.64 0.2952 1.5 1 1.66 6.326 2.8 1 1.58 0.2638 1.8 1 1.68 5.961 1.5 1 1.43 0.2384 2.1 ! 1.68 5.848 7.0 2 1.44 0.1968 2.2 1 1.66 5.740 2.9 2 1.48 0.1631 2.4 1 1,65 5.438 1.1 2 1.55 0.1362 2.7 1 1,70 5.253 8.7 3 1.58 0.1215 3.7 1 1.74 5.166 7.6 3 1.58 0.1181 4.0 1 1.75 4.881 1.0 2 1.61 0.1159 4.3 1 1.75 4.626 2.7 2 1.62 0.1097 5.0 I 1.72 4.592 2.8 2 1.61 0.1016 5.8 1 1.67 4.492 1.4 2 1.61 0.0925 6.7 1 1.58 4.428 1.1 2 1.61 0.0800 7.7 I 1.37 4.217 1.0 2 1.63 0.0785 7.7 1 1.33 3.948 1.3 2 1.64 0.0588 6.2 1 0.963 3.668 2.1 2 1.65 0.0449 3.8 1 0.812 3.463 3.5 2 1.65 0.0415 3.1 1 0.802 3.246 5.6 2 1.65 0.0312 1.4 1 0.850 3.009 7.5 2 1.61 0,0207 4.9 2 0.920

was then used to obtain n over the entire E'k(E') range from 0.025 to 1000/xm. To perform n(E)- 1 = _2f~ dE', (9) 7'1" the integration in Eq. (8), the k 0Q values 134 KHARE ET AL.

i i iiii111 i I iiilll I i [ ,lllll I i I II11111 i i iiiilll

2.0

1.6

1.2

0.8 ~:...... ::...... ,, ...~ ...... I ..... I I I I :;::: ...... , ,,...I i ......

0.i / / k 0.01

0,001

0.000] I I ttJmi I l llllnl I I lllllll i l llllnl I I I TM 0,01 0,I l fo i00 iOO0 k(#m)

FIG. 4. Real, n, and imaginary, k, values of the complex refraclive index of Titan tholins as a function of wavelength. Determined by Kramers-Kronig analysis (--). Also shown are n values obtained independently by ellipsometery (©), from interference patterns in transmission (0), from Brewster angle measurements (A), and from measured R and k values (...). The dotted curve (...) between 0.2 and 0.4 ~m is discussed in the text. The straight dashed line (---) at ,X <. 0.025 #m is used for the Kramers-Kronig analysis and is also discussed in the text. were extrapolated at ?~ < 0.025 /~m, as ond Kramers-Kronig analysis, the mea- shown by the straight dashed line in Fig. 4. sured k values from 0.15 to 70/~m and from The short wavelength cutoff for this Kra- 100 to 1000 p~m, and the interpolated values mers-Kronig integration was adjusted to between 70 and 100 p.m were used and the make the calculated values of n agree with structure present in R from 0.2 to 0.4 /xm the measured values in the 0.4- to 2.3-/zm- was lost. The values of n, calculated using wavelength region. The n values obtained Eq. (7) and the measured values of R and k, from the Kramers-Kronig analysis of k are are plotted as the dotted curve in Fig. 4 for exhibited as the solid curve in Fig. 4. How- 0.2 p~m _< X -< 0.4/~m. Also shown in Fig. 4 ever, in the wavelength region from 0.2 to are the independently measured values of 0.4 /~m, the measured reflectance (Fig. 3) n. Real parts of refractive indices from the shows a structure which was not repro- Kramers-Kronig analysis are higher than duced by the Kramers-Kronig analyses. the measured values in the 300- to 1000-/~m- This is because the first Kramers-Kronig wavelength range, probably because of analysis, performed to obtain k values from density differences between the tholin films 0.025 to 0.15 /~m, did not reproduce the and pellets. small values of k obtained experimentally The normal-incidence Fresnel reflec- from 0.2 to 0.4 k~m. In performing the sec- tance, R, has been calculated from Eq. (7) OPTICAL CONSTANTS OF ORGANIC THOLINS 135

TABLE II INSTRUMENTS AND TECHNIQUES EMPLOYED IN THE MEASUREMENT OF OPTICAL CONSTANTS OF TITAN THOLINS MADE FROM 0.9 Nff0.1 CH4

Wavelength Instrument Measurement (~m) Normal Real part of Imaginary part of reflectance, refractive refractive index, k, R, from index, n from transmittance reflectance

0.025 -9.13 McPherson Model 247 grazing incidence monochromator X 0.13-0.15 Seya-Namioka monochromator X 0.15-0.4 Seya-Namioka monochromator X X 0.4-0.5 Seya-Namioka monochromator Brewster angle X 0.5-0.6 Seya-Namioka monochromator Brewster angle X 0.4-2.0 Oak Ridge National Laboratory Model HP-E-60/70 ellipsometer Ellipsometry 0.4-2.5 Cary Model 14PM spectrometer X 1.0-2.3 Cary Model 14PM spectrometer Interferometry 2.5-40 Perkin-Elmer Model 521 infrared spectrometer X 2.0-70 Nicolet Series 8000 Fourier transform infrared spectrometer X 100-300 Nicolet Series 8000 Fourier transform infrared spectrometer X 300-1000 Nicolet Series 8000 Fourier transform infrared spectrometer Interferometry X for wavelengths above 0.4/xm, and the val- lengths. But it is difficult to make films ues plotted in Fig. 5. thick enough to guarantee negligible trans- To summarize, in the Kramers-Kronig mittance at h > 0.4/zm, and at h < 0.025 method, n and k can in principle be deter- /~m. Alternatively, k can in principle be mined by measuring R over all wave- measured over all wavelengths by measuring T, and then n extracted by a separate

t0 ...... i ...... i ...... i ...... i ...... Kramers-Kronig analysis. But, even for 0.9N2/0.1 CH4 THOLIN our thinnest films, T is too small to measure at ~ < 0.15 /xm. Therefore, we have adopted a mixed strategy, the individual measurements for which are summarized in Table II. 0.(7t / The experimental uncertainties are diffi- / cult to evaluate precisely, but are estimated / to be -+30% in the measured k values and 000t /, ...... i ...... i ...... i ...... ~ ...... 001 0.t 1.0 t0 100 t000 +3% in the measured n values. The uncer- X (/~m) tainty in k arises primarily from sample-to- FIG. 5. Normal incidence reflectance calculated us- sample variations. Additional sources of ing the n and k values shown in Fig. 4. The dashed line uncertainty in both n and k are small errors is due to values of k extrapolated to shorter wavelengths, as shown by the dashed line in Fig. 4. The in the measurement of reflectance. dotted curve between 0.2 and 0.4/zm is the measured So far as we know, values ofn and k have reflectivity plotted in Fig. 3. (See text for details.) never before been obtained over so wide a 136 KHARE ET AL. wavelength range for any complex organic The experimental program described material. The values of the imaginary part here may be useful in measuring, with of the refractive index, k, of the tholin vary higher accuracy and broader wavelength from almost 1 near I000 A to about 3 × 10 4 coverage than has been achieved hereto- near 1.5/xm. The slope of k between these fore, the optical constants of soots and two wavelengths--responsible, inciden- smokes, important for testing the nuclear tally, for the red color of the tholin and, winter hypothesis (Turco, et al., 1983). presumably, of Titan as well--is due to the superposition of a number of electronic transitions. In the infrared, a number of vi- ACKNOWLEDGMENTS brational transitions are readily identifiable, We are grateful to J. Osantowski of Goddard Space including C--H near 3/zm and C~N near Flight Center, NASA, for use of the Nicolet Fourier 4.6 txm. The infrared spectrum of such tho- transform infrared spectrometer, to J. Heaney and K. Stewart for experimental assistance, and to W. R. lins are discussed elsewhere (cf. Sagan et Thompson for helpful discussions. This research was at., 1984). From 1000 A to microwave fre- supported in part by the National Aeronautics and quencies, the real part of the refractive in- Space Administration, Grant NGR 33-010-101, and by dex n of the tholin stays approximately con- the Office of Health and Environmental Research, stant, wandering sluggishly from -~1.6 to U.S. Department of Energy, under Contract W-7405- eng-26 with the Union Carbide Corporation. ~2.2, and reflecting in muted fashion the sharper gradients in k, as expected from classical dispersion theoroY (e.g., Christy, REFERENCES 1972). Shortward of 1000 A, n falls precipi- ABELES, F. (1950). La determination de I'indice et de tously to values below 0.8. l'epaisseur des couches minces transparentes. J. Estimates of the real part, n, of the com- Phys. Radium 11, 310-314. plex refractive index of the Titan aerosol CHRISTY, R. W. (1972). Classical theory of optical dis- layer, made from ground-based photomet- persion. Amer. J. Phys. 40, 1403-1419. HALL, J. F., AND W. F. C. FERGUSON (1955). Optical ric observations (Rages and Pollack, 1980); properties of cadmium sulfide and zinc sulfide from from Pioneer 11 polarimetric observations 0.6 microns to 14 microns. J. Opt. Soc. Amer. 45, (Tomasko and Smith, 1982); and from Voy- 714-718. ager photometric observations (Rages and HANEL, R., B. CONRATH, F. M. FLASAR, V. KUNDE, Pollack, 1981, 1983, and private communi- W. MAGUIRE, J. PEARL, J. PIRRAGLIA, R. SAM- UELSON, L. HERATH, M. ALLISON, D. CRUIK- cation), all suggest values for visible wave- SHANK, D. GAUTIER, P. GIERASCH, L. HORN, R. lengths of 1.65 -+ 0.1. In the same wave- KOPPANY, AND C. PONNAMPERUMA ( 1981 ). Infrared length range, our tholins exhibit values n = observations of the Saturnian system from Voyager 1.65 -+ 0.05, in excellent agreement with I. Science (Washington, D.C.) 212, 192-200. ground-based and spacecraft observations. INAGAKI, T., E. T. ARAKAWA, R. HAMM, AND M. W. WILLIAMS (1977). Optical properties of polystyrene The same Titan observations suggest (see from the near-infrared to the x-ray region and con- also Rages et al., 1983" Podolak, 1984) val- vergence of optical sum rules. Phys. Rev. B 15, ues ofk in the visible between 0.1 and 0.01, 3243-3253. which again corresponds to the values pre- INAGAKI, T., L. C. EMERSON, E. T. ARAKAWA, AND sented here for our CH4/N2 tholin. Radia- M. W. WILLIAMS (1976). Optical properties of solid Na and Li between 0.6 and 3.8 eV. Phys. Rev. B 13, tive transfer calculations, to be published 2305-2313. separately, reveal a close agreement be- JAHODA, F. C. (1957). Fundamental absorption of bar- tween the observed spectrum of Titan in ium oxide from its reflectivity spectrum. Phys. Retd. the ultraviolet, visible, and near infrared, 107, 1261-1265. and the spectrum calculated assuming that KAYE, G. W. C., AND T. H. LABY (1966). Tables qf Physical and Chemical Constants. pp. 83-86. Wi- the Titan aerosol is composed in significant ley, New York. part of the organic tholin whose complex KHARE, B. N., AND C. SAGAN (1973). Red clouds in refractive index is given in Fig. 4. reducing atmospheres. Icarus 20, 311-321. OPTICAL CONSTANTS OF ORGANIC THOLINS 137

KHARE, B. N., AND C. SAGAN (1975). Cyclic octa- SAGAS, C. (1973). The greenhouse of Titan. Icarus 18, tomic sulfur: A possible infrared and visible chro- 649-656. mophore in the clouds of Jupiter. Science (Washing- SAGAS, C. (1974). Organic Chemistry in the Atmo- ton, D.C.) 189, 722-723. sphere: The Atmosphere of Titan (D. M. Hunten, KHARE, B. N., C. SAGAN, S. SHRADER, AND E. T. Ed.), NASA Special Publication SP-340, pp. 134- ARAKAWA (1982). Molecular analysis of tholins pro- 142. U.S. Govt. Printing Office, Washington, D.C. duced under simulated Titan conditions. Bull. Amer. SAGAS, C., AND B. N. KHARE (1979). Tholins: Or- Astron. Soc. 14, 714. ganic chemistry of interstellar grains and gas. Na- KHARE, B.N., C. SAGAS, E. T. ARAKAWA, H. OGINO, ture (London) 277, 102-107. T. O. WILLINGHAM, AND B. NAGY (1983). Amino SAGAS, C., AND B. N. KHARE (1981). The organic acid analysis of Titan tholins. Bull. Amer. Astron. clouds of Titan. Bull. Amer. Astron. Soc. 13, 701. Soc. 15, 843. SAGAS, C., AND B. N. KHARE (1982). The organic KHARE, B. N., C. SAGAS, AND J. GRADIE (1981). Re- clouds of Titan. Orig. Life 12, 280. flection spectra of simulated Titan organic clouds. SAGAS, C., AND W. R. THOMPSON (1984). Production Bull. Amer. Astron. Soc. 13, 701. and condensation of organic gases in the atmosphere KUNDE, V. G., A. C. AIKEN, R. A. HANEL, D. E. of Titan. Icarus 59, 133-161. JENNINGS, W. C. MAGUIRE, AND R. E. SAMUELSON SAGAS, C., B. N. KHARE, AND J. LEWIS (1984). Or- (1981). C4H2, HC3N and C2N2 in Titan's atmo- ganic matter in the Saturn system. In Saturn (M. S. sphere. Nature (London) 292, 686-688. Matthews and T. Gehrels, Eds.). Univ. of Arizona MAGUIRE, W. C., R. A. HANEL, D. E. JENNINGS, V. Press, Tucson, in press. G. KUNDE, AND R. E. SAMUELSON (1981). C3Hs and SMITH, B., L. SODERBLOM, R. BEEBE, J. BOYCE, G. C3H4 in Titan's atmosphere. Nature (London) 292, BRIGGS, A. BUNKER, S. A. COLLINS, C. J. HANSEN, 683-686. T. V. JOHNSON, J. L. MITCHELL, R. J. TERRILE, M. PODOLAK, M. (1984). Are the polarization data consis- CARR, A. F. COOK II, J. CuzzI, J. B. POLLACK, G. tent with constant flux models of Titan's atmo- E. DANIELSON, A. INGERSOLL, M. E. DAVIES, G. E. sphere? Icarus 58, 325-329. HUNT, H. MASURSKY, E. SHOEMAKER, D. MORRI- RAGES, K., AND J. B. POLLACK (1980). Titan aerosols: SON, T. OWEN, C. SAGAN, J. VEVERKA, R. STROM, Optical properties and vertical distribution. Icarus AND V. E. SUOMI (1981). Encounter with Saturn: 41, 119-130. Voyager 1 imaging science results. Science (Wash- RAGES, K., AND J. B. POLLACK (1981). High phase ington, D.C.) 212, 163-190. angle Voyager images of Titan's main aerosol layer. SMITH, G. R., D. F. STROBEL, A. L. BROADFOOT, B. Bull. Amer. Astron. Soc. 13, 703. R. SANDEL, D. E. SHEMANSKY,AND J. B. HOLBERG RAGES, K., AND J. B. POLLACK (1982). Aerosol extinc- (1982). Titan's upper atmosphere: composition and tion profiles in Titan's atmosphere. Bull. Amer. As- temperature from the EUV solar occultation results. tron. Soc. 14, 714. J. Geophys. Res. 87, 1351-1360. RAGES, K., AND J. B. POLLACK (1983). Vertical distri- TOMASKO, M. G., AND P. H. SMITH (1982). Photome- bution of scattering hazes in Titan's upper atmo- try and polarimetry of Titan: Pioneer 11 observa- sphere. Icarus 55, 50-62. tions and their implications for aerosol properties. RAGES, K., J. B. POLLACK, AND O. B. TOON (1983). Icarus 51, 65-95. Radiative-convective equilibrium temperature pro- TURCO, R. P., O. B. TOON, T. P. ACKERMAN, J. B. files in Titan's atmosphere. Bull. Amer. Astron. Soc. POLLACK, AND C. SAGAS (1983). Nuclear Winter: 15, 841. Global consequences of multiple nuclear explo- SAGAS, C. (1971). The solar system beyond Mars: An sions. Science (Washington, D.C.) 222, 1283- exobiological survey. Space Sci. Rev. 11, 73-112. 1292.